kern/41014: media auto strangenesses with if_wm and ukphy

[spz%NetBSD.org@localhost]

>Number:         41014
>Category:       kern
>Synopsis:       media auto strangenesses with if_wm and ukphy
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Mar 14 12:15:00 +0000 2009
>Originator:     S.P.Zeidler
>Release:        NetBSD 5.0_RC2 20090314
>Organization:
        TNF
>Environment:
System: NetBSD babylon5.netbsd.org 5.0_RC2 NetBSD 5.0_RC2 (NBFTP) #10: Sat Mar 
14 09:48:40 UTC 2009 
spz@ADMIN:/chroots/netbsd-5/src/sys/arch/amd64/compile/obj/NBFTP amd64
Architecture: x86_64
Machine: amd64
>Description:

the console log rained:
Mar 14 10:47:44 babylon5 /netbsd: wm0: MDIC read timed out: phy 1 reg 5
Mar 14 10:48:10 babylon5 /netbsd: wm0: MDIC write timed out: phy 1 reg 22
Mar 14 10:48:11 babylon5 /netbsd: wm0: MDIC read timed out: phy 1 reg 0
Mar 14 10:48:15 babylon5 /netbsd: wm0: MDIC write timed out: phy 1 reg 22
times many on me.

the server says:
        367# ifconfig wm0
wm0: flags=8c43<UP,BROADCAST,RUNNING,OACTIVE,SIMPLEX,MULTICAST> mtu 1500
        
capabilities=7ff80<TSO4,IP4CSUM_Rx,IP4CSUM_Tx,TCP4CSUM_Rx,TCP4CSUM_Tx,UDP4CSUM_Rx,UDP4CSUM_Tx,TCP6CSUM_Rx,TCP6CSUM_Tx,UDP6CSUM_Rx,UDP6CSUM_Tx,TSO6>
        
enabled=7ff80<TSO4,IP4CSUM_Rx,IP4CSUM_Tx,TCP4CSUM_Rx,TCP4CSUM_Tx,UDP4CSUM_Rx,UDP4CSUM_Tx,TCP6CSUM_Rx,TCP6CSUM_Tx,UDP6CSUM_Rx,UDP6CSUM_Tx,TSO6>
        address: 00:30:48:31:43:f2
        media: Ethernet autoselect (100baseTX 
full-duplex,flowcontrol,rxpause,txpause)
        status: active
the switch says:
  Full-duplex, 10Mb/s, media type is 10/100/1000BaseTX

well at least they agree on full-duplex :7

I tell the switch we want 1000BaseTX (speed 1000). The interface goes down,
the interface goes up, the connection works. The server says:

wm0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
        
capabilities=7ff80<TSO4,IP4CSUM_Rx,IP4CSUM_Tx,TCP4CSUM_Rx,TCP4CSUM_Tx,UDP4CSUM_Rx,UDP4CSUM_Tx,TCP6CSUM_Rx,TCP6CSUM_Tx,UDP6CSUM_Rx,UDP6CSUM_Tx,TSO6>
        
enabled=7ff80<TSO4,IP4CSUM_Rx,IP4CSUM_Tx,TCP4CSUM_Rx,TCP4CSUM_Tx,UDP4CSUM_Rx,UDP4CSUM_Tx,TCP6CSUM_Rx,TCP6CSUM_Tx,UDP6CSUM_Rx,UDP6CSUM_Tx,TSO6>
        address: 00:30:48:31:43:f2
        media: Ethernet autoselect (none)
        status: active

media none, hu?

the switch says:
  Full-duplex, 1000Mb/s, media type is 10/100/1000BaseTX

the error messages on console stopped, and as far as I can see the system
is as happy as a clam now.

This problem did not manifest with the previously running kernel
(if_wm/ukphy, netbsd-5 of 20090222) or earlier (which of course may have been
plain dumb luck, seeing as there were no suspicious pullups to either
if_wm nor ukphy in the respective timeframe as far as I'm aware).

>How-To-Repeat:
        
>Fix: